rattus 09bcbddfcf ModelPatcherDynamic: Force load all non-comfy weights (#12739) ... * model_management: Remove non-comfy dynamic _v caster * Force pre-load non-comfy weights to GPU in ModelPatcherDynamic Non-comfy weights may expect to be pre-cast to the target device without in-model casting. Previously they were allocated in the vbar with _v which required the _v fault path in cast_to. Instead, back up the original CPU weight and move it directly to GPU at load time.		2026-03-03 08:50:33 -08:00
..
audio_encoders	…
cldm	…
comfy_types	feat: add gradient-slider display mode for FLOAT inputs (#12536)	2026-02-20 22:52:32 -08:00
extra_samplers	…
image_encoders	…
k_diffusion	ace15: Use dynamic_vram friendly trange (#12409)	2026-02-11 14:53:42 -05:00
ldm	Feat: z-image pixel space (model still training atm) (#12709)	2026-03-02 19:43:47 -05:00
sd1_tokenizer	…
t2i_adapter	…
taesd	…
text_encoders	fix(ace15): handle missing lm_metadata in memory estimation during checkpoint export #12669 (#12686)	2026-02-28 01:18:40 -05:00
weight_adapter	MPDynamic: force load flux img_in weight (Fixes flux1 canny+depth lora crash) (#12446)	2026-02-15 20:30:09 -05:00
cli_args.py	model_mangament: make dynamic --disable-smart-memory work (#12724)	2026-03-01 19:18:56 -08:00
clip_config_bigg.json	…
clip_model.py	…
clip_vision.py	…
clip_vision_config_g.json	…
clip_vision_config_h.json	…
clip_vision_config_vitl.json	…
clip_vision_config_vitl_336.json	…
clip_vision_config_vitl_336_llava.json	…
clip_vision_siglip2_base_naflex.json	…
clip_vision_siglip_384.json	…
clip_vision_siglip_512.json	…
conds.py	Cleanups to the last PR. (#12646)	2026-02-26 01:30:31 -05:00
context_windows.py	fix: handle substep sigmas in context window set_step (#12719)	2026-03-01 09:38:30 -08:00
controlnet.py	Add working Qwen 2512 ControlNet (Fun ControlNet) support (#12359)	2026-02-13 22:23:52 -05:00
diffusers_convert.py	…
diffusers_load.py	…
float.py	…
gligen.py	…
hooks.py	…
latent_formats.py	Feat: z-image pixel space (model still training atm) (#12709)	2026-03-02 19:43:47 -05:00
lora.py	feat(ace-step): add ACE-Step 1.5 lycoris key alias mapping for LoKR #12638 (#12665)	2026-02-26 18:19:19 -05:00
lora_convert.py	Use torch RMSNorm for flux models and refactor hunyuan video code. (#12432)	2026-02-13 15:35:13 -05:00
memory_management.py	comfy-aimdo 0.2 - Improved pytorch allocator integration (#12557)	2026-02-21 10:52:57 -08:00
model_base.py	Feat: z-image pixel space (model still training atm) (#12709)	2026-03-02 19:43:47 -05:00
model_detection.py	Feat: z-image pixel space (model still training atm) (#12709)	2026-03-02 19:43:47 -05:00
model_management.py	ModelPatcherDynamic: Force load all non-comfy weights (#12739)	2026-03-03 08:50:33 -08:00
model_patcher.py	ModelPatcherDynamic: Force load all non-comfy weights (#12739)	2026-03-03 08:50:33 -08:00
model_sampling.py	initial FlowRVS support (#12637)	2026-02-25 23:38:46 -05:00
nested_tensor.py	…
ops.py	ops: implement lora requanting for non QuantizedTensor fp8 (#12668)	2026-02-27 19:05:51 -05:00
options.py	…
patcher_extension.py	…
pinned_memory.py	fix pinning with model defined dtype (#12208)	2026-02-01 08:42:32 -08:00
pixel_space_convert.py	…
quant_ops.py	…
rmsnorm.py	Remove code to support RMSNorm on old pytorch. (#12499)	2026-02-16 20:09:24 -05:00
sample.py	…
sampler_helpers.py	Disable dynamic_vram when weight hooks applied (#12653)	2026-02-28 16:50:18 -05:00
samplers.py	Disable dynamic_vram when weight hooks applied (#12653)	2026-02-28 16:50:18 -05:00
sd.py	Disable dynamic_vram when weight hooks applied (#12653)	2026-02-28 16:50:18 -05:00
sd1_clip.py	Force min length 1 when tokenizing for text generation. (#12538)	2026-02-19 22:57:44 -05:00
sd1_clip_config.json	…
sdxl_clip.py	…
supported_models.py	Feat: z-image pixel space (model still training atm) (#12709)	2026-03-02 19:43:47 -05:00
supported_models_base.py	…
utils.py	Fix Aimdo fallback on probe to not use zero-copy sft (#12634)	2026-02-25 16:49:48 -05:00
windows.py	…